Public access radio that connects community members to one another and the world
Play Live Radio
Next Up:
0:00
0:00
0:00 0:00
Available On Air Stations
Join KDNK for the Solstice Songwriter Session in the KDNK Gardens on June 20th at 6pm

A.I. and the movies

ANDREW LIMBONG, HOST:

The new Netflix true crime documentary "What Jennifer Did" tells the story of a young Canadian woman involved in a 2010 murder-for-hire scheme targeting her parents.

(SOUNDBITE OF DOCUMENTARY, "WHAT JENNIFER DID")

UNIDENTIFIED PERSON: It was a situation that I could never have imagined.

LIMBONG: There's been some criticism, not because the movie is overly grisly or anything, but because people suspect the filmmakers used generative AI in the doc. That's just one of a handful of stories about AI in moviemaking that popped up this week. Here with us to break them all down is NPR's Chloe Veltman. Hey, Chloe.

CHLOE VELTMAN, BYLINE: Hi there, Andrew.

LIMBONG: All right, so how did this documentary, "What Jennifer Did," allegedly use AI? And why are people so mad about it?

VELTMAN: Well, social media has gone bonkers about what looks like the use of AI-generated or manipulated photographs of the young woman at the center of the film, Jennifer Pan. She's currently serving a life sentence for a kill-for-hire attack that killed her mother and severely injured her father. So the images appear around the half-hour mark in the movie, when a high school friend, Nam Yin (ph), describes Pan's personality.

(SOUNDBITE OF DOCUMENTARY, "WHAT JENNIFER DID")

NAM YIN: Jennifer, you know, was bubbly, happy, confident and very genuine.

VELTMAN: His words were accompanied by a series of images showing a young and ebullient Pan. But if you look closely, there are these weird glitches. For instance, Pan's fingers - they're all mangled. And human hands can be a sign that an image has been manipulated because they're really tricky to generate and edit. So people calling the alleged image manipulation disgraceful on social media - to quote one Reddit user, "it devalues and betrays the whole idea of true crime, not to mention mocking the victims by distorting the facts."

LIMBONG: What have we heard from Netflix and the filmmakers?

VELTMAN: Well, NPR reached out to Netflix for comment, and they didn't want to talk about it. But the film's executive producer, Jeremy Grimaldi, denied the use of AI in an interview he gave with The Toronto Star published on Friday. However, he did admit to the use of photo editing software.

LIMBONG: What exactly is the fuss here? Let's just, like, take him at his word, and, you know, say he did just use some photo editing software. Or even if he did use AI, is this not, like, a like a tom-ay-to (ph), tom-ah-to (ph) type thing?

VELTMAN: Well, the public is just starting to become much more aware of suspicious-looking images in general owing to the proliferation of AI, Andrew. I mean, just think about the angry reactions a few weeks ago to British royal Kate Middleton's doctored family photo, right? So I spoke with documentary production veteran Jennifer Petrucelli about why people's reactions to these types of things are so strong.

JENNIFER PETRUCELLI: People don't like to be deceived, generally. You know, in the cases where there hasn't been transparency, it is where you've seen pushback.

VELTMAN: Petrucelli is part of a group called the Archival Producers Alliance that went public this week with a draft of guidelines for how documentary filmmakers can work ethically with AI.

PETRUCELLI: Because when you sit down to watch a documentary, you're making certain assumptions about what you're seeing. A big one is that what you're seeing is true.

VELTMAN: And Petrucelli's group and others are urging filmmakers to pay special attention to accuracy and transparency when using AI and documentary films.

LIMBONG: All right. Like we said before, that was just one in a couple of different AI in movies stories. What else you got?

PETRUCELLI: Yeah. There are a few things going on, Andrew. So the production company A24 has been catching a lot of heat for the poster campaign it unveiled on social media to promote its latest blockbuster, "Civil War." The series of glossy images show these apocalyptic scenes of U.S. cities under siege - so a building on fire in Chicago, a destroyed Miami street and armed soldiers emerging from a lake in San Francisco and that sort of thing.

LIMBONG: Which are, you know, the sorts of scenes that are typical for kind of, like, fictional apocalyptic movies, right? How is that any different from, you know, what we've been seeing for years with CGI?

VELTMAN: Well, people are getting upset for a couple of reasons. So, one, none of the scenes in the posters are in the actual movie, which is causing a bit of confusion.

LIMBONG: Oof.

VELTMAN: Yes, a little bit of annoyance. And, two, many folks are unhappy about the fact the images may have been made with AI. People are asking why this massive, expensive Hollywood production is using machines to create PR campaigns rather than human artists. That's one big thing that's come up. And also, once again, people don't like being taken for a ride with images that just look weird and off. It's that uncanny valley feeling that no one likes.

That being said, it's hard to know for sure if the images really were created using AI. NPR reached out to A24 to ask, but no response. Most news outlets that have picked this up are referencing a story in The Hollywood Reporter. It quotes an unnamed source as evidence of the use of AI, which is admittedly a bit suspect. But, you know, regardless, there's certainly some wonky-looking stuff in those posters, Andrew. I mean, for instance, the location of the buildings in the Chicago image are a bit off. And the Miami poster has a weird car with three doors.

LIMBONG: (Laughter) AI hasn't quite figured out it's either two or four generally, right?

VELTMAN: Yeah. That's right. Hands and doors, apparently very tricky.

LIMBONG: All right. So I spent a couple of minutes this week being pretty enthralled by a trailer for a fake James Bond movie. Can you tell me a bit more about it?

VELTMAN: Yeah. This is a fun one. So a video creator by the name of KH Studio made this speculative trailer using AI for a Bond film that will probably never get made. It stars Henry Cavill as 007, speaking in an American accent, a first for this quintessentially British special agent.

(SOUNDBITE OF ARCHIVED RECORDING)

AI-GENERATED VOICE #1: (As 007) What makes you think this is my first time?

VELTMAN: And Margot Robbie as the love interest.

(SOUNDBITE OF ARCHIVED RECORDING)

AI-GENERATED VOICE #2: (As character) Hello, James Bond.

VELTMAN: It's gone viral. Millions of people have viewed the video since it was released a few days ago, and they're reacting positively.

LIMBONG: Well, what's the difference here? I mean, why were people mad about, like, the other things but cool with this?

VELTMAN: It's an issue of transparency, basically. I mean, for one thing, the video maker is completely transparent. There's a clear statement on YouTube accompanying the trailer saying it was made with AI purely for entertainment. Oh, but also, there's been a lot of rumors lately about who will get to play the next James Bond. Many fans would like Henry Cavill to snag the part. So I think people are enjoying imagining the actor has already been cast. And, of course, Margot Robbie's presence in the movie is a lovely act of wishful thinking.

LIMBONG: That's NPR's Chloe Veltman. Chloe, thanks so much.

VELTMAN: Such a pleasure, Andrew. Transcript provided by NPR, Copyright NPR.

NPR transcripts are created on a rush deadline by an NPR contractor. This text may not be in its final form and may be updated or revised in the future. Accuracy and availability may vary. The authoritative record of NPR’s programming is the audio record.

Chloe Veltman
Chloe Veltman is a correspondent on NPR's Culture Desk.